An Intelligent Approach for Power Quality Event Classification

The hybrid power systems which means that integration of conventional energy sources with utility grid creates the challenges in terms of power system stability and Power Quality (PQ). The reasons for PQ events are huge due to increase in use of sensitive loads, power electronic devices, nonlinear loads and incorporation of nonconventional energy sources to the power grid. Due to these PQ disturbances, the equipments are not operated properly and may lead to the risk of black out. Hence it is very essential task to the power system engineers to recognize the PQ disturbances by studying the distorted waveforms. This paper suggests an intelligent technique based on EMD-HT-ELM (Empirical Mode Decomposition-Hilbert transform-Extreme Learning Machine) for feature extraction (energy, standard deviation (std) of amplitude and phase) and classification of the PQ disturbances. Various PQ disturbances such as sag, swell, harmonics with sag and swell, spike, notch and harmonics have been generated in MATLAB/SIMULINK environment. Simulation results have been observed in terms of EMD-HT-ELM scheme which gives superior classification accuracy (97.6%) as compared to schemes by EMD-HT-PNN (probabilistic neural network) and EMD-HT-SVM (support vector machine).